[PATCH] memcg_write_event_control(): fix a user-triggerable oops

Al Viro posted 1 patch 2 months, 3 weeks ago
[PATCH] memcg_write_event_control(): fix a user-triggerable oops
Posted by Al Viro 2 months, 3 weeks ago
We are *not* guaranteed that anything past the terminating NUL
is mapped (let alone initialized with anything sane).
    
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
diff --git a/mm/memcontrol-v1.c b/mm/memcontrol-v1.c
index 2aeea4d8bf8e..417c96f2da28 100644
--- a/mm/memcontrol-v1.c
+++ b/mm/memcontrol-v1.c
@@ -1842,9 +1842,12 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
 	buf = endp + 1;
 
 	cfd = simple_strtoul(buf, &endp, 10);
-	if ((*endp != ' ') && (*endp != '\0'))
+	if (*endp == '\0')
+		buf = endp;
+	else if (*endp == ' ')
+		buf = endp + 1;
+	else
 		return -EINVAL;
-	buf = endp + 1;
 
 	event = kzalloc(sizeof(*event), GFP_KERNEL);
 	if (!event)
Re: [PATCH] memcg_write_event_control(): fix a user-triggerable oops
Posted by Michal Hocko 2 months, 3 weeks ago
On Fri 26-07-24 06:43:57, Al Viro wrote:
> We are *not* guaranteed that anything past the terminating NUL
> is mapped (let alone initialized with anything sane).
>     

Fixes: 0dea116876ee ("cgroup: implement eventfd-based generic API for notifications")

> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

Acked-by: Michal Hocko <mhocko@suse.com>

> ---
> diff --git a/mm/memcontrol-v1.c b/mm/memcontrol-v1.c
> index 2aeea4d8bf8e..417c96f2da28 100644
> --- a/mm/memcontrol-v1.c
> +++ b/mm/memcontrol-v1.c
> @@ -1842,9 +1842,12 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
>  	buf = endp + 1;
>  
>  	cfd = simple_strtoul(buf, &endp, 10);
> -	if ((*endp != ' ') && (*endp != '\0'))
> +	if (*endp == '\0')
> +		buf = endp;
> +	else if (*endp == ' ')
> +		buf = endp + 1;
> +	else
>  		return -EINVAL;
> -	buf = endp + 1;
>  
>  	event = kzalloc(sizeof(*event), GFP_KERNEL);
>  	if (!event)

-- 
Michal Hocko
SUSE Labs
Re: [PATCH] memcg_write_event_control(): fix a user-triggerable oops
Posted by Michal Hocko 2 months, 3 weeks ago
On Fri 26-07-24 09:06:59, Michal Hocko wrote:
> On Fri 26-07-24 06:43:57, Al Viro wrote:
> > We are *not* guaranteed that anything past the terminating NUL
> > is mapped (let alone initialized with anything sane).
> >     
> 
> Fixes: 0dea116876ee ("cgroup: implement eventfd-based generic API for notifications")
> 
> > Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> 
> Acked-by: Michal Hocko <mhocko@suse.com>

Btw. this should be
Cc: stable

> 
> > ---
> > diff --git a/mm/memcontrol-v1.c b/mm/memcontrol-v1.c
> > index 2aeea4d8bf8e..417c96f2da28 100644
> > --- a/mm/memcontrol-v1.c
> > +++ b/mm/memcontrol-v1.c
> > @@ -1842,9 +1842,12 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
> >  	buf = endp + 1;
> >  
> >  	cfd = simple_strtoul(buf, &endp, 10);
> > -	if ((*endp != ' ') && (*endp != '\0'))
> > +	if (*endp == '\0')
> > +		buf = endp;
> > +	else if (*endp == ' ')
> > +		buf = endp + 1;
> > +	else
> >  		return -EINVAL;
> > -	buf = endp + 1;
> >  
> >  	event = kzalloc(sizeof(*event), GFP_KERNEL);
> >  	if (!event)
> 
> -- 
> Michal Hocko
> SUSE Labs

-- 
Michal Hocko
SUSE Labs
Re: [PATCH] memcg_write_event_control(): fix a user-triggerable oops
Posted by Al Viro 2 months, 3 weeks ago
On Fri, Jul 26, 2024 at 09:13:00AM +0200, Michal Hocko wrote:
> On Fri 26-07-24 09:06:59, Michal Hocko wrote:
> > On Fri 26-07-24 06:43:57, Al Viro wrote:
> > > We are *not* guaranteed that anything past the terminating NUL
> > > is mapped (let alone initialized with anything sane).
> > >     
> > 
> > Fixes: 0dea116876ee ("cgroup: implement eventfd-based generic API for notifications")
> > 
> > > Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> > 
> > Acked-by: Michal Hocko <mhocko@suse.com>
> 
> Btw. this should be
> Cc: stable

Point, except that for -stable it needs to be applied to mm/memcontrol.c
instead...
Re: [PATCH] memcg_write_event_control(): fix a user-triggerable oops
Posted by Michal Hocko 2 months, 3 weeks ago
On Fri 26-07-24 08:48:04, Al Viro wrote:
> On Fri, Jul 26, 2024 at 09:13:00AM +0200, Michal Hocko wrote:
> > On Fri 26-07-24 09:06:59, Michal Hocko wrote:
> > > On Fri 26-07-24 06:43:57, Al Viro wrote:
> > > > We are *not* guaranteed that anything past the terminating NUL
> > > > is mapped (let alone initialized with anything sane).
> > > >     
> > > 
> > > Fixes: 0dea116876ee ("cgroup: implement eventfd-based generic API for notifications")
> > > 
> > > > Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> > > 
> > > Acked-by: Michal Hocko <mhocko@suse.com>
> > 
> > Btw. this should be
> > Cc: stable
> 
> Point, except that for -stable it needs to be applied to mm/memcontrol.c
> instead...

Correct. And if anybody wants to backport to pre 3.14 then to
kernel/cgroup.c

-- 
Michal Hocko
SUSE Labs